Hacker News Reader

Comment by druiid | Hacker News Reader

Eh, AUFS isn't that odd these days. I built a chroot build utility for an open-source project on the back of it and it's super easy to work with. Haven't dealt with it in a couple of years, but I can't imagine it somehow got all kinds of crazy difficult to deal with or implement.

View on HN

I haven't been able to find a clear reason as to why AUFS seems to have lost support from major distros - although the fact that it requires patching VFS in the kernel and isn't "just a module" doesn't help.

Yeah. That part is true. I think you are probably onto the reason there. For RHEL distributions at least all I've seen included are the fuse based unioning filesystems which is sad.

Don't think that's it. OverlayFS was accepted and also requires patching VFS.

If overlayFS is in-tree, then it's a possibility to be backported by Red Hat. From the discussions on the GitHub issue tracker, there's a definite problem with custom kernels vs RHEL supported w/ contract.

My understanding is that it's been unable to get into the Linux kernel because it's unstable? I agree it's a great tool if it works :-)

Docker maintainer here. I love aufs, it's been awesome for us at dotCloud. It is definitely stable in production (at least it has been for us). I really really wish we didn't have to remove it as the default. And I don't really know why it's not being merged, exactly. I sure there are good reasons but they are not immediately obvious to mere mortals like me.

In any case, aufs will come back as an optional storage backend in a future version. It still has a few advantages over device-mapper - increased memory sharing for example.

Just my 2c.

Interesting! What was the response when you tried to get it merged into the kernel?

I had nothing to do with the effort with getting it merged. But Junjiro Okajima (the author of aufs) seems to have tried a few times, and obviously it hasn't worked. At the same my observation as a "real world implementor" is that it works great and no obvious alternative is available.

Again I'm not an expert in "kernel politics" (you could even say I know nothing about them).

Union filesystems in general have had a storied and troubled past in the Linux kernel's political landscape. Valerie Aurora (formerly Valerie Henson), a former redhat kernel developer who worked on Union filesystems tried quite valiantly (no pun intended!) to get the unionfs patchset merged and ultimately left redhat to pursue personal things.

Somewhat recently there was some traction with a version of her patches that looked somewhat close to being merged, but it might have also been dropped.

http://lwn.net/Articles/482779/

Reply on news.ycombinator.com